368 



MEMOIRS NATIONAL ACADEMY OF SCIENCES. 



[Vol. XV, 



The tests of Investigation 3 and also examination a were given to 70 unselected literate 

 men. The correlation of composite scores of the five tests with examination a was 0.88. The 

 correlation of the separate tests with examination a was as follows: Maze, 0.456; picture se- 

 quence, 0.766; series completion, 0.724; line length, 0.483; pictorial analogies, 0.772. 



This group of tests was more original in conception and more satisfactory in its results 

 than any other series devised for the purpose. Although only two of the five tests were included 

 in the final beta scale, the ideas which they embodied proved very helpful when the beta trial 

 series was constructed in January, 1918. 



Section 2. — The beta trial series. 



Experiments in the four camps had shown that the Stenquist test was unsatisfactory for 

 the purpose for which it was intended; also that other types of performance tests, requiring 

 no material except pencd and paper, which could be given expeditiously and scored by stencil, 

 were feasible. Accordingly, in the latter part of December, 1917, arrangement was made by 

 the staff of the Division of Psychology to construct a group examination for foreigners and 

 illiterates. Capt. Wm. S. Foster, who had been in charge at Camp Devens, directed the work. 

 W. S. Hunter (who worked for a short time only), K. T. Waugh, A. S. Otis, C. R. Brown, R. H. 

 Wheeler, and W. P. Tomlinson assisted. 



The purposes and requirements of this examination (beta) were formulated as follows 

 by the group : 



The (beta) examination is intended primarily for the group testing of foreigners and of 

 others too nearly illiterate hi English to do themselves justice hi examination alpha. 



All oral and written instructions, except a few simple phrases required in maintaining 

 order and securing attention, should be avoided. Instructions should be gi\ 7 en, and with the 

 assistance of a trained orderly, in the form of four demonstrations at the beginning of each test. 

 Gestures and pantomine should be used where such procedures are serviceable and can be 

 standardized. 



Relation of beta examination to alpha examination: (a) Beta examination is designed 

 for men making a score below a certain point in the literacy test. It is, therefore, a substitute 

 for alpha primarily for lower grade men. (b) Scores in beta should be comparable with scores 

 hi alpha and the former should be standardized hi terms of the latter, (c) Letter grades A, B, C, 

 D and E should be given on the basis of beta score just as for alpha, (d) The beta tests should 

 be patterned after the alpha tests as far as possible. 



The beta examination should test a variety of intellectual processes. The separate tests 

 should range in difficulty from every easy to difficult. 



The scale should meet the following requirements: (a) It should be suitable for admin- 

 istration to groups of 25 to 100 men. (b) The examination time should be not more than 

 one hour, (c) No material should be required except pencil and paper, (d) All responses 

 should be made by drawing line, checking, or crossing out, with the possible exception of writing 

 numbers, (e) Each test should be scored by stencils. (/) Duplicate "forms'' should be 

 possible. 



In order to reduce individual examining as much as practicable, there should be some 

 very easy tests with high diagnostic ralue. The instructions should be such as low grade 

 men can readily comprehend. 



With the above criteria in view, the following tests were selected for trial: 



1. Maze. 



2. Form recognition. 



3. Number recognition. 



4. Dot imitation. 



5. Pictorial completion. 



6. Cube analysis. 



7. Line comparison. 



8. Picture sequence. 



9. Digit, symbol. 



10. Spot pattern. 



1 1 . Picture analogies. 



12. Geometrical construction 

 (Form board representation). 



13. Series completion. 



14. Picture situation. 



15. Memory for designs. 



An idea of the nature of each test may be gained from the procedure as described here- 

 inafter, and from the test blanks which are reproduced in Part I, (pp. 235 to 258). 



